Transformers Stop Thinking Too Early, and a Tiny LoRA Fixes It Paper • 2609.36585 • Published 15 days ago • 72
The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation Paper • 2609.36484 • Published 15 days ago • 558
False Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents Paper • 2609.39102 • Published 14 days ago • 654
Edge0/Audio8-ASR-Infinite Automatic Speech Recognition • 4B • Updated 20 days ago • 41.6k • 2.45k